Methodology — AI Governance Monitor

The AI Governance Monitor applies a consistent research standard across all 16 modules every week. This page documents that standard.

Source Hierarchy

All items are sourced according to a three-tier hierarchy. Tier 1 is always preferred; lower tiers are only used when no Tier 1 source exists.

Signal Standard

An item is included if and only if:

1. It is new within the 7-day reporting window
2. A senior professional in the target audience would want to know about it
3. It has a primary source link
4. It is not already covered by another item in the same report

There are no arbitrary item caps. If a module has 10 signal-quality items, all 10 are included.

Forensic Filters

Four filters are applied across every issue, regardless of module:

Science Drill-Down (M06): Every week: AlphaFold release notes, OpenAI Preparedness trigger disclosures, Anthropic RSP threshold triggers, DeepMind programme updates. Any protein structure volume update or RSP threshold trigger is flagged regardless of mainstream coverage.

Energy Wall (M03): Secondary sweep beyond standard >$50M investment rounds: Physics-ML, Thermodynamic Computing, AI Energy Infrastructure. Flags structural shift from model investment to compute/energy sovereignty.

Ciyuan / Standards Vacuum (M05 & M09): China: monitor for state-level AI token commodity framing (“Ciyuan” / 词元). EU: flag any amendment that widens the gap between when a legal obligation applies and when the harmonised compliance standard is available.

AISI-to-Lab Pipeline (M15): Priority weekly scan of senior departures from UK AISI, US AISI, Canadian CAISI, or EU AI Office into frontier labs (OpenAI, Anthropic, DeepMind, xAI, Meta AI, Mistral, Cohere). Also flagged: reverse direction (lab → regulator).

Module Guide

Accountability Friction

For every item in M11 (Ethics & Accountability), the report explicitly assesses the gap between a lab’s stated commitments and its observed actions. This is not optional — it is a required field for every corporate and lab item in that module.

Asymmetric Signal

Every module item that warrants it includes an “Asymmetric Signal” — a non-obvious 12-month implication that a mainstream observer would miss. This is the core editorial value of the monitor: not what happened, but what it means for the trajectory.

Reporting Window

Each issue covers Monday to Sunday of the preceding week. Published Friday at 18:00 CEST.

Cross-Monitor Signals

Each issue of the AI Governance Monitor includes a Cross-Monitor Signals section in the dashboard. This section identifies overlaps, reinforcing signals, and conflicts with the other monitors on asym-intel.info.

Scope

The AI Governance Monitor scans the public dashboards and briefs of all other monitors published on asym-intel.info before each weekly compilation. The monitor set is treated generically — new monitors are included automatically as they appear, without any change to this methodology.

What constitutes a cross-monitor flag

A flag is raised when developments in another monitor’s domain materially affect, or are materially affected by, AI governance dynamics. The flag names the other monitor, describes the linkage in 2–4 sentences, and states the AI Governance Monitor’s specific perspective.

Update rules

Cross-monitor flags follow the same data lifecycle rules as all other persistent entries. A structural linkage is not re-described merely because a week has passed. Flags are updated only when the nature, direction, or evidence base of the linkage materially changes. Closed flags are archived, not deleted.

No-signal case

If no material cross-monitor signals exist in a given period, the section states this explicitly. The section is present in every issue without exception.

Sources

Only public dashboards, briefs, and methodology pages from asym-intel.info are used as sources for cross-monitor flags. Internal operational details of other monitors are never referenced.

Data Lifecycle

The monitor builds a cumulative intelligence picture, not a transient news feed. Entries are not deleted because time has passed.

Persistent data

The following remains visible until something material changes:

• Policy positions, doctrines, legal frameworks, military postures
• Baseline deviations and confirmed risk ratings
• Ongoing conflicts, campaigns, or active monitoring flags (Standards Vacuum, Ciyuan, AISI Pipeline)
• The EU AI Act 7-layer tracker (M09) and Concentration Index (M14)

Transient data

Single announcements, one-off events, dated statements, and tactical incidents may be summarised or rolled into higher-level series once their implications are captured. They are never silently deleted — they are archived as closed episodes.

Update rules

An entry is updated only if:

• New data materially changes the assessment (substance, direction, or level of concern)
• Confidence improves (e.g. Probable → Confirmed) or degrades
• Source quality changes (key claims now supported by higher-tier sources)

An entry is not updated merely because a week has passed, or to republish identical findings under a new date.

Version history

Every persistent entry carries a version history: timestamp, what changed, and why. Past assessments are never silently overwritten. When a persistent state closes (e.g. a risk rating drops, a campaign ends), it is logged as a closed episode with an end date and final assessment.

Persistent state file

The current state of all persistent entries is maintained in data/persistent-state.json. Each weekly issue reads this file as its baseline and updates only entries that meet the update rules above. The weekly brief references this evolving knowledge base — it does not recreate it from scratch.

Editor

Peter Howitt · asym-intel.info · Gibraltar